Vehicle identification is required today in several applications, such as toll roads, parking lots, electronic surveillance including, e.g., security and law enforcement applications, etc. Presently, conventional vehicle identification is based on the reading of registration strings on vehicle license plates. There are known methods for License Plate Recognition (LPR), also known as Automatic Number Plate recognition (ANPR). LPR includes automatically reading a vehicle registration string off of images of a license plate captured by video cameras and/or still cameras. A registration string may include alphanumeric characters and/or other signs.
For example, U.S. Pat. No. 6,339,651 to Tian et al. describes a method and system for recognizing characters on surfaces where an alphanumeric identification code (“ID” for short) may be present such as a license plate. Tian's system is particularly adapted for situations where visual distortions can occur, and utilizes a highly robust method for recognizing the characters of an ID. Multiple character recovery schemes are applied to account for a variety of conditions to ensure high accuracy in identifying the ID. Special considerations are given to recognizing the ID as a whole and not just the individual characters.
However, PCT Patent Application Publication No. WO/2005/041071 to Lawida et al., for example, realizes that it may be possible to remove a license plate from one vehicle and attach it to another vehicle. It also possible to copy, counterfeit or spoof the license plate and attach it to other vehicles. Consequently, vehicle identification based solely on LPR is not truly vehicle identification, rather only recognition of the associated object that is intended to be used in conjunction with a vehicle. Accordingly, a need exists for an improved solution for vehicle identification. For example, in addition to license plate recognition, it may be desirable to recognize additional parameters which identify the vehicle, such as make and model, special vehicle types such as handicapped, and other attributes.
US Patent Application Publication No. 2005/0270178 to Ioli, describes a system for parking enforcement that allows vehicles to be identified and tracked without operator involvement. The system includes a meter system that generates image data of a vehicle by creating an array of pixel data in a predetermined field of view that includes a vehicle identification tag and facial imaging. An enforcement and tracking system receives the image data and generates a vehicle license number, a vehicle tag identification number and a facial image from the image data, such as by analyzing the image data to identify the vehicle license number, vehicle tag identification number and facial image based on the expected location of the license tag, identification tag and field of view image data characteristics of the license tag, facial image or other suitable data. From the image data acquired, monitoring of parking spaces is performed and violation citations or notices are generated for errant vehicles in parking locations as well as notification to law enforcement and homeland security agencies of vehicles and facial images identified as being on a watch list.
US Patent Application Publication No. 2008/0285804 to Sefton, describes a system for identifying the state of issuance of a license plate. The system analyzes various design characteristics of a vehicle license plate, including character size, placement and color, to identify the state of issuance of the plate. In some embodiments, the system uses spectral properties of light reflected from a vehicle license plate to determine spectral frequency bands having the best contrast between characters on the plate and the background of the plate. For example, red characters against a white background exhibit high contrast levels at wavelengths of about 420 nm to about 595 nm. Green characters against a white background exhibit high contrast levels at wavelengths of about 600 nm to about 750 nm. Blue characters against a white background exhibit high contrast levels at wavelengths of about 550 nm to about 750 nm. Thus, spectral characteristics in combination with other design-related characteristics of a license plate may be used to identify the state of origin of the plate. Once the state of origin is identified, origin-specific syntax matching may be used to enhance optical character recognition routines.
Referring again to the Lawida application, Lawida references a method for vehicle recognition using a plurality of metrics from one or more vehicle sensor(s). Lawida suggests creating a multimetric vehicle identification profile comprising at least two metrics and matching the obtained multimetric profile against a plurality of stored vehicle sensor recordings. However, Lawida does not teach how to process data obtained from the vehicle sensors and therefore the Lawida's referenced method remains unknown and cannot be implemented.
Several publications describe image recognition systems pertaining to finding correspondences between two images of the same scene or object. The article “SURF: Speeded Up Robust Features”, by Bay H., Tuytelaars T., and Van Gool L. (ECCV 2006), describes detecting ‘interest points’ at distinctive locations in the image, such as corners, blobs, and T-junctions. “Interest points” are also known as “key-points” or “feature points”. The neighbourhood of every interest point is represented by a vector named “descriptor vector”. Then, the descriptor vectors are matched between different images. Detecting the interest points is done by a module named “detector”, while representing the interest point's neighbourhood by a feature vector is performed by a “descriptor”.
The following articles describe other existing interest points detectors and descriptors:    Harris, C., Stephens, M.: A combined corner and edge detector; Proceedings of the Alvey Vision Conference. (1988) 147-151. The methods described in this article are hereinafter referred to as the “Harris corner detector”.    Mikolajczyk, K., Schmid, C.: Indexing based on scale invariant interest points. In: ICCV. Volume 1. (2001) 525-531. The methods described in this article are hereinafter referred to as “Modified Harris methods”.    Lowe, D.: Object recognition from local scale-invariant features. In: ICCV. (1999). The methods described in this article are hereinafter referred to as “SIFT”, for Scale Invariant Feature Transform.
Further to discussing interest points, attention is drawn now to image processing, wherein a color histogram is a representation of the distribution of colors in an image. For digital images, a color histogram represents the number of pixels that have colors in each of a fixed list of color ranges that span the image's color space, or the set of all possible colors. That is, while considering images of an object, it can be appreciated that the object can be modeled using color histograms.
Color histograms can be built from images in various color spaces, whether RGB, RG chromaticity, HSV or any other color space of any dimension. Basically, a histogram of an image is produced first by discretization of the colors in the image into a number of bins, and counting the number of image pixels in each bin.
F. Mindru, T. Tuytelaars, L. Van Gool and T. Moons provide a paper entitled “Moment invariants for recognition under changing viewpoint and illumination”, (Computer Vision and Image Understanding, 94(1-3):3-27, 2004). In this paper they teach that when objects are viewed under different angles and different lighting conditions, their image displays photometric and geometric changes. This means that the image colors are different, and geometric deformations like scaling, rotation, and skewing have to be taken into account. A variety of approaches exist to the problem of identifying the presence of the same object under such photometric and/or geometric changes. One way of proceeding is to estimate the transformations and compensate for their effects. An alternative is deriving invariant features, that is deriving features that do not change under a given set of transformations. The main advantage of using invariants is that they eliminate expensive parameter estimation steps like camera and light source calibration or color constancy algorithms, as well as the need for normalization steps against the transformations involved.
In the article “Color indexing”, Michel Swain and Dana Ballard (International Journal of Computer Vision, 7(1), 1991) use color histograms of model objects. They explain that the image colors that are transformed to a common discrete color are usefully thought of as being in the same 3D histogram bin centered at that color. Histograms are invariant to translation and rotation about the viewing axis, and change only slowly under change of angle of view, change in scale, and occlusion.
Those versed in the art would appreciate that different histograms exist. Some examples are RGB histogram, opponent histogram, hue histogram and RG histogram.
In the HSV color space, it is known that the hue becomes unstable around the grey axis. To this end, J. van de Weijer, T. Gevers, and A. Bagdanov, in their article “Boosting color saliency in image feature detection” (IEEE Trans. Pattern Analysis and Machine Intell., 28(1):150-156, 2006.) relating to hue histograms, discuss salient feature detection, whose aim is to find distinctive local events in images. Salient features are generally determined from the local differential structure of images. They focus on the shape-saliency of the local neighborhood. The majority of these detectors are luminance-based, which has the disadvantage that the distinctiveness of the local color information is completely ignored in determining salient image features. To fully exploit the possibilities of salient point detection in color images, color distinctiveness should be taken into account in addition to shape distinctiveness.
Each of the above cited references are herein expressly incorporated by reference in their respective entireties.
There is a need in the art, thus, for a mechanism for collecting information relating to identity parameters of a vehicle.